refactor validator to wrap Codex API

[kelsey-wong]

Key Info

• Notion ticket
• Priority: normal

What changed?

Remove redundant validator logic from client library and instead wrap the Codex validator API. The backend now handles the following logic in the projects.validate endpoint:

• Run all evals against TrustworthyRAG
• Apply eval thresholds, if provided, to determine if the RAG app's generated response is good or bad
  • If the response is determined to be bad, then query Codex project for an expert answer that matches the query

What do you want the reviewer(s) to focus on?

Elias: Please confirm if the ML team needs/wants any changes to the client library functionality

Angela: General code review (test coverage, naming, format, etc.)

---

Checklist

• Did you link the GitHub issue?
• Did you follow deployment steps or bump the version if needed?
• Did you add/update tests?
• What QA did you do?
  • Tested...